DNA methylation

In mammalians, DNA methylation occurs at the 5′- position of cytosine residues, mainly within CpG dinucleotides, 60–80 % of which are methylated within the promoter regions of genes [24]. Methylation of CpG dinucleotides within the promoter regions leads to the silencing of transcription process, mediated by modifications in the condensation status of the chromatin. The process of DNA methylation is catalyzed by enzymes known as DNA methyltransferases (DNMTs), which can be classified in “de novo” DNMTs (which methylate specific chromosomal sequences during early embryogenesis), and maintenance methyltransferases (DNMT1), faithfully restoring the methylation patterns after each DNA replication cycle [26]. The process of DNA methylation is closely related with gametogenesis, since primordial germ cells (PGCs), when entering the developing gonad, undergo a process of deep decrease of DNA methylation, which will be subsequently restored in the prenatal life in males and during post-natal follicle development in females (Fig. 1) [27].

Cytosine methylation can occur also in non-CpG sites (CpA > CpT > CpC), although the significance of these variants is still unknown. Methylation at non-CpG sites was previously considered to be largely present in the brain, embryonic stem cells (ESCs), induced pluripotent stem cells, and in oocytes [28–32]. However, Ichiyanagi et al. observed methylation at non-CpG sites in male germ cells as well, demonstrating that the level of the non-CpG methylation is higher in prospermatogonia and decreases along with mitotic division. In addition, this study also suggested the absence of a template-dependent mechanism for copying non-CpG methylation in prospermatogonia [33].

Another kind of epigenetic DNA modification is represented by 5-hydroxymethylcytosine (5hmC), an intermediate of DNA demethylation with important regulatory functions in various biological and pathological processes in the mammalian genome such as transcriptional regulation, DNA methylation regulation, and tumorigenesis [34]. Gan et al. showed that the highly ordered 5hmC alterations are critical for the differentiation of spermatogenic cells in the mouse [35]. Recently, Wang et al. for the first time compared 5-hydroxymethylcytosine profiles in normal, abnormal, and globozoospermia sperms, identifying 6664, 9029, and 6318 genes containing 5hmC, respectively [36]. In addition, since some 5hmC-containing genes are significantly involved in spermatogenesis, sperm motility, and morphology, the authors suggested that the 5hmC distribution differences may contribute to the sperm phenotype [36].

DNA methylation was originally investigated by Southern blot and methylation-sensitive restriction endonuclease digestion followed by PCR amplification [37]. Nowadays, gene-specific techniques are available including bisulfite sequencing, combined bisulfite restriction analysis (COBRA) [38], MethyLight [39], and bisulfite pyrosequencing [40], all based on bisulfite conversion of cytosine residue to uracil, leaving 5-methylcytosine residue unaffected [41], detectable by DNA sequencing. All the abovementioned techniques are sensitive, specific, and relatively inexpensive but not suitable for analysis of the whole genome, which includes about 28 million CpGs.

The next-generation sequencing (NGS) approach represents a new powerful tool for the study of human DNA methylome [42]. The NGS DNA methylation procedures are based on three main steps: digestion of genomic DNA with methyl-sensitive restriction enzymes, affinity-based enrichment of methylated DNA fragments, and chemical conversion methods [43, 44].

Another large-scale approach is represented by whole-genome bisulfite sequencing (WGBS), which is able to analyze DNA methylation profiles of whole genomes at single-base resolution [45]. Recently, several studies have reported WGBS accurate data on methylome both of human embryonic stem cells [46] and induced pluripotent stem cells [47].

Histone modifications

Modifications of histone tails represent other epigenetic chromatin marks critical for transcriptional regulation. Post-translational modifications of histone tails include methylation, acetylation, phosphorylation, ubiquitination, ribosylation, and sumoylation. Histone marks are dynamic process, since histone modifications can be easily induced and removed by a wide range of enzymes [48]. The most relevant histone change is represented by acetylation at lysine residues on the amino-terminal tail domains, whose correct levels are maintained by the combined action of two enzymes known as histone acetyl transferase (HAT) and histone deacetylase (HDAC). Generally, acetylation reduces the affinity of histones for DNA, making genes functionally active; on the other hand, histone deacetylation leads to chromatin condensation, making genes transcriptionally inactive. Conversely, histone methylation is a key regulator for both activation and inactivation of transcription. For example, lysine 4 of histone H3 (H3-K4) methylation is linked to gene expression, while H3K9 and H3K27 di- and tri-methylation is associated with gene silencing [49, 50].

Histone H3T is the testis-specific H3 variant in mammals. Tachiwana et al. showed that H3T nucleosomes can be assembled by Nap2 chaperone and that this structure is significantly unstable as compared to the conventional H3.1-containing nucleosomes [51, 52]. Furthermore, these authors suggested that the unstable H3T-containing nucleosome structure can influence the chromatin re-organization essential for spermatogenesis [52].

Histone modifications were originally investigated by Western blot with specific antibodies to modified histones, but nowadays, they are mainly studied by mass-spectrometry-based proteomic technologies [53]. Recently, NGS technologies have been developed able to analyze whole-genome histone modifications. These techniques are based on chromatin immunoprecipitation followed by sequencing maps of the genome-wide binding pattern of chromatin-associated proteins, which includes modified histones (ChIP-seq) [44, 54].

Small non-coding RNAs

A third mechanism of epigenetic regulation of gene activity is represented by small RNAs and large intergenic non-coding RNAs (lincRNAs). Small RNAs, not encoding proteins, are located in the nucleus of spermatozoa and play an essential role as epigenetic modifiers both in recognizing and preserving the DNA that remains bound to histones during the transition to protamine in spermiogenesis and in early embryonic development [55]. On the other hand, lincRNAs act at the level of chromatin allowing to select the histone modification enzymes, such as in the case of lincRNAs HOTAIR which acts as a scaffold for PRC2 and LSD1 enzymes to regulate lysine 27 methylation and lysine 4 demethylation in H3 histone [56].

Small non-coding RNAs are generally identified by RT-PCR analysis, in situ hybridization, or small RNA sequencing studies. Furthermore, the recent development of microarray technologies has permitted the global analysis of spermatozoal microRNAs (miRNAs) evidencing different expression profiles between fertile and infertile men [57, 58].

The process of spermatogenesis

The formation of a mature sperm requires different processes, namely (1) mitotic proliferation of spermatogonia; (2) meiotic divisions; and (3) morphological differentiation of sperm precursors (spermiogenesis), leading to the generation of highly specialized cells characterized by the presence of a head, an intermediate portion, and a flagellum. Such a specific organization of the male germ cells is necessary to allow sperms to traverse a potentially hostile female reproductive tract, penetrate the cumulus oophorus and the zona pellucida, penetrate the oocyte, and finally complete multiple post-penetration events [62, 63]. During fetal life, spermatogenesis begins in the wall of the seminiferous tubules from undifferentiated diploid cells known as spermatogonia, which undergo several mitotic divisions in order to increase the pool of available precursors of germ cells. At puberty, some spermatogonia are transformed in type I spermatocytes, which undergo the first meiotic division producing haploid type II spermatocytes. A second meiotic division occurs in these cells, originating haploid spermatids. The last phase of spermatogenesis is represented by spermiogenesis, characterized by a morphological and structural transformation complex process of the round spermatid. This step, occurring without further cell division, leads to the production of mature sperm, characterized by the differentiation of the flagellum and the acrosome, essential prerequisites for sperm motility and fertilization capacity.

Histone–protamine replacement as the main epigenetic change in sperms

In addition to the typical morphology and motility, sperms are characterized also by a highly organized chromatin structure. In fact, sperm chromatin during spermiogenesis undergoes further condensation, due to the replacement of 90–95 % of the histones with one or more sperm-specific basic proteins, known as protamines [64]. This modification induces the formation of disulfide bonds (SS) that confer extreme stability to the core of the sperm nucleus, producing a number of relevant effects, such as improvement of sperm motility, protection from oxidative stress and toxic agents present within female reproductive tract, and block of the transcriptional activity of the sperm DNA [65]. The complex mechanism of histone–protamine transition is a finely regulated multi-step process. In the first step, the histones in round spermatids are replaced by a heterogeneous group of nuclear proteins (transition proteins (TP)), as the result of histone hyperacetylation [66] (Fig. 1). The second step takes place in elongating spermatids, determining the replacement of TP1 and TP2 with protamines [62, 67] (Fig. 1). Protamines have different functions: they allow the compaction of the nucleus ensuring the genetic integrity of the sperm and play an important role in epigenetic imprinting [62]. Mature spermatid nuclei present two types of protamines: the P1 protamine and the P2 family of protamines, constituted by P2 (the most abundant), P3, and P4 members. P1/P2 ratio appears to be critical for male fertility [68, 69]. In fact, the P1/P2 ratio, which in fertile males is close to 1 (range 0.8–1.2), is altered in infertile patients [69, 70]. Patients with a P1/P2 ratio <0.8 present inadequate DNA condensation and important alterations in sperm parameters, such as motility, counts, and structure [68–71]. Moreover, Aoki et al. demonstrated that low P1/P2 ratios are also associated with an increased DNA fragmentation, which is also inversely correlated with global sperm P1 and P2 concentrations, suggesting a protective role of the protamines against sperm DNA damage [72]. There is also evidence that subfertility can be correlated with an excess of protamine P2 precursors (pre-P2), determined by an alteration of the process leading to the mature protamine P2 formation [68, 71, 73, 74].

DNA methylation and histone modifications during spermatogenesis

Various and specific epigenetic marks are required during male gametogenesis for proper maturation of gametes. In fact, before meiosis, the first epigenetic events take place in the form of progressive demethylation–remethylation of DNA. During meiosis, DNMT3A, DNMT3B, and cofactor DNMT3L activity regulates the levels of de novo DNA methylation, completing this process after birth at the stage of pachytene spermatocyte [75]. Subsequently, the methylation profile is maintained by DNMT1 activity. In addition to the above-described processes, also histone modifications (methylation and acetylation) occur, which modify DNA accessibility to transcription factors (Fig. 1). In fact, specific enzymes such as histone methyltransferase (HMT) and histone demethylase (HDM) regulate lysine 9 of histone H3 (H3-K9) and lysine 4 of histone H3 (H3-K4) methylation patterns. Generally, histone H3-K9 methylation is high in meiosis but is removed at the end of this process, promoting gene activation, whereas histone H3-K4 methylation, which decreases during meiosis, is associated to DNA silencing [50] (Fig. 1). In addition, during spermatogenesis, several enzymes, such as HAT and HDAC, regulate the processes of acetylation and deacetylation of H3 and H4 lysine residues. During spermiogenesis, hyperacetylation of H4 plays a crucial role for correct histone to protamine transition and allows nucleosome disassembly in elongating spermatids [66, 76] (Fig. 1).

Epigenetic alterations and spermatogenesis disruption

The above-described epigenetic marks in germ line genes play a key role in the proper spermatogenesis processes, and several studies have demonstrated that aberrant epigenetic modification of genes expressed in the testes are associated with male infertility. Navarro-Costa et al. for the first time hypothesized that one of the factors of male gametogenic defects could be represented by epigenetic alterations of specific genes, evidencing increased methylation defects of the germ line regulator DAZL gene in different quality-fractioned sperm populations of oligoasthenoteratozoospermic (OAT) patients as compared to normozoospermic (NZ) men [77]. On the contrary, no variation in the methylation state of the DAZL gene promoter between NZ e OAT men was observed. This study also highlighted the existence of homogeneous DAZL methylation levels when comparing the normal sperm-enriched fractions of NZ men. This evidence has been subsequently confirmed on a larger scale by Krausz et al. who detected no differences in the DNA methylation status of several genes in the different sperm subpopulations of normozoospermic individuals [78]. Subsequently, other studies have identified more genes whose epigenetic modifications are related to alteration of both semen parameters and fertility maintenance. Hammoud et al., by analyzing seven imprinted loci (LIT1, MEST, SNRPN, PLAGL1, PEG3, H19, and IGF2), correlated alterations in the DNA methylation pattern of oligozoospermic patients with abnormal protamine levels [70]. Interestingly, alterations in the mRNA levels of the genes involved in the histone–protamine transition have been evidenced also by studies carried out by investigating the global testis transcriptome of normal and oligozoospermic patients by microarray analysis [79]. Other studies have demonstrated that DNA hypermethylation of the promoter of several genes (such as MTHFR, PAX8, NTF3, SFN, HRAS, RASGFR1, GTL2, PLAG1, D1RAS3, MEST, KCNQ1, LT1, SNRPN and others) plays a critical role in male infertility, being associated to alterations of sperm concentration, motility, and morphology, while hypomethylation of the IGF2/H19 imprinting control region 1 (ICR1) has been detected in patients with low concentration and sperm motility as compared to normozoospermic controls [70, 80–86].

Due to the increased knowledge about the epigenetic alterations occurring in sperm DNA of infertile patients, it has become clear that specific errors in the processes of epigenetic control may occur during each stage of spermatogenesis, adversely affecting male fertility and embryonic development [49]. In fact, epigenetic alterations occurring in mitosis can affect the expression of specific genes involved in the first steps of spermatogenesis, leading to a decreased efficiency of the process. When the meiotic stage is involved, these alterations can induce double-strand breaks or chromosomal non-disjunction. Finally, during spermiogenesis, epigenetic alterations can involve histone-to-protamine transition and histone removal and degradation, inducing protamine replacement errors [49]. Taken together, all these evidences suggest that different features of male infertility, such as alterations in sperm count or morphology, DNA fragmentation, chromosomal aneuploidies, and alterations in the chromatin package, could be all related to epigenetic mechanisms occurring at different stages of spermatogenesis.

Data from ART in animal models

Early studies carried out in mice had showed that alterations affecting the development and growth of the fetuses were linked to ovulation induction, manipulation of eggs, or embryo culture in vitro [116–119]. A confirmation came from Khosla et al. showed that preimplantation mouse embryos cultured in the presence of serum can change the expression and methylation of several imprinted genes (such as H19, IGF2, GRB10, and GRB7) and that these aberrant epigenetic modifications lead to abnormal fetal growth in ART animals [120].

Moving to different animal models, Young et al. evidenced the presence of the “large offspring syndrome” (LOS) (large size at birth, increased birth weight, breathing difficulties, reluctance to suckle, and sudden perinatal death) in sheep and cattle derived from cultured embryo [121]. Subsequently, the pathogenesis of LOS was suggested to be associated with epigenetic abnormalities, leading to loss of imprinting and over-expression of IGF2 receptor gene [122, 123]. Factors in the ART procedures, triggering these imprinting errors, were not clearly identified, but the onset of the syndrome appeared to be dependent by in vitro culture conditions [122, 123].

Data from ART in human

In human, a larger prevalence of syndromes related to imprinting alteration, particularly Beckwith-Wiedemann and Angelman syndromes, has been reported in children born after ART when compared to non-ART children [124–126]. Interestingly, the phenotype of Beckwith-Wiedemann syndrome, an overgrowth condition characterized by large size at birth, macroglossia, and visceromegaly, closely remembers the previously described LOS in cows and sheep generated by in vitro fertilization, suggesting a common mechanism of origin of these conditions. On the other hand, also a disproportionate number of low birth weight cases has been observed in ART children [127].

The association between ART and altered DNA epigenetic profiling has been demonstrated also by studies based on the analysis of the methylation status of CpG sites in the promoters of 700 genes of placenta and cord blood obtained from children conceived in vitro and in vivo [128]. This analysis showed hypomethylation of most CpG sites in the placenta and hypermethylation of most CpG sites in cord blood in the group of children conceived by in vitro fertilization as compared to naturally conceived children. Interestingly, the genes showing different expressions in the two groups appeared to be involved in chronic metabolic disorders including obesity, type II diabetes, and high blood pressure [128]. These data are in agreement with other studies reporting an increased risk of disturbs in body fat composition, changes in blood pressure, and increase in the late infancy growth velocity in children generated by ART procedures as compared to control children [129–131]. In addition, other studies showed a high risk of obesity or type II diabetes in adult life in children generated by ART [132]. These results strongly suggest that the effect of ART procedures could be manifested not only at birth, but also in late infancy or even in adult life.

As in animal models, the in vitro culture conditions have been suggested as a main cause of epigenetic defects in the offspring generated by ART also in human [133]. In addition, it has been proposed that the low birth weight observed after ART, as well as in small for gestational age and very premature children, is the result of an unfavorable embryonic, fetal, or neonatal environment, involving also epigenetic mechanisms, potentially related to metabolic alterations in late childhood [134]. However, several evidences have suggested that a crucial role could be also played by epigenetic defects in the sperms used in ART protocols. In fact, several authors suggested that DNA methylation changes at imprinted loci are inherited from the sperm of men with oligozoospermia [82, 135, 136]. This could suggest that the increase of DNA methylation variations in ART depends, at least in part, on the presence of epigenetic defects of male gamete. In this view, some authors have proposed the possible usefulness of analyzing the imprinting methylation status in the routine sperm examination for ART treatment [136]. Moreover, it has been hypothesized that the use of sperms with an abnormal P1/P2 ratio or defects of histone–protamine may be responsible of imprinting diseases in the offspring conceived with ART [50]. Unfortunately, this argument remains poorly understood, as demonstrated by other studies reporting that epigenetic abnormalities detected in sperms of oligozoospermic patients do not appear to be associated with ART outcome, suggesting that further studies are required in order to shed light on the relationship between sperm epimutation and alterations in children generated by ART [137].

Paternal exposure to toxins or ionizing radiation

Great attention has been devoted to the effects of paternal exposures to environmental toxins or low-dose ionizing radiation, and of paternal lifestyle [145]. Several studies had previously demonstrated the presence of a strong association between paternal occupational exposures to chemicals and harmful health outcomes in the offspring. Feychting at al. demonstrated an increased risk of nervous system tumors related to paternal occupational exposure to pesticides and of leukemia related to woodwork by fathers [146]. Reid et al. evidenced the presence of high exposure to exhausts by paternal grandmothers of children with acute lymphoblastic leukemia [147]. However, many of these conditions are likely related to the presence of mutations in sperm DNA, thus representing a genetic, rather than epigenetic, mechanism. Is there any evidence supporting the presence of epigenetic mechanism driving the effects to the offspring of the paternal exposure to chemicals? Once again, the most relevant data in support of this hypothesis come from studies on animal models, showing that male exposure to pesticides or other harmful chemicals can be responsible for defects in the gametes and abnormal development of the offspring mainly via altered DNA methylation patterns in the germ line [148, 149]. Anway et al. evidenced that a transient embryonic exposure to the endocrine disruptor vinclozolin during gonadal sex determination in rats produced several diseases affecting the prostate, kidney, immune system, testis, as well as different cancers in the subsequent generations, suggesting a potential transgenerational effect [148]. Similar results were obtained by Guerrero-Bosagna et al., who showed that transient exposure of the F0 generation gestating female to vinclozolin during gonadal sex determination caused adult onset disease in the F3 generation male and female mice [149].

Also, ionizing radiations have been recently invoked as a risk factor for alterations of DNA methylation. These radiations trigger a series of processes on the cells as genotoxic alterations including DNA breaks, but the actual mechanism leading to a transgenerational effect is still poorly understood. Dubrova et al. suggested an epigenetic mechanism of transmission of the radiation-exposure signal through sperm, likely involving DNA methylation and affecting DNA repair processes [150]. These authors suggested that the persistence of instability into the germ line of unexposed offspring of irradiated mice could be responsible of mosaicism in germ cells, a well-known mechanism in the origin of human genetic disorders [150].

More recently, it has been suggested that a crucial role in transgenerational radiation effects, such as genomic and epigenomic instability, could be played by the Piwi-interacting RNAs (piRNA) pathway, involved in the maintenance of genomic stability by facilitating DNA methylation of transposable elements and also implicated in other epigenetic alterations affecting a variety of cellular regulation processes [151]. Another experiment on animal models supports transgenerational epigenetic changes as a result of parental exposure to genotoxic stressors, as irradiation, nutrition, and intake of anti-androgen compounds [152]. For example, it has been demonstrated that treatment with the anti-androgen compound vinclozolin on female mice induces epigenetic effects in the sperm of their offspring as compared to controls [153]. This study highlighted an increased methylation of the differentially methylated domains (DMDs) of maternal PEG1, PEG3, and SNRPN genes and decreased methylation of paternal H19 and GTL2 genes.

Paternal diet

One of the most intriguing topics in the field of epigenetic modifications of the germ line is represented by the influence played by the paternal diet on gametogenesis. The first evidences of this association came from animal models. Carone et al. demonstrated that male mice fed with low-protein diet generated an offspring showing an increased expression of genes involved in the synthesis of lipids and cholesterol, as compared to the offspring of control male mice fed with a normal diet. Based on these results, authors suggested that cholesterol and lipid metabolism in an offspring can be strongly affected by paternal diet [154]. This study was carried out by a whole-genome characterization of cytosine methylation patterns and RNA content in sperm obtained from mice submitted to low-protein or caloric restriction diets and controls. Authors detected similar cytosine methylation patterns in all three conditions, thus suggesting that the sperm epigenome is largely unaffected by these diets and that changes in relatively few loci can have profound effects in the developing animal [154]. However, in a more recent study, Radford et al. demonstrated that in utero undernourishment perturbs the adult sperm methylome, suggesting that alterations in gamete methylation could induce alterations in chromatin architecture, transcriptional networks differentiation, or tissue structure, and in turn is able to contribute to the intergenerational transmission of environmentally induced diseases [155].

In another animal study, Ng et al. showed the presence of pancreatic alterations, with early onset impaired insulin secretion and glucose tolerance worsening with time, in the female offspring of male mice fed with a high-fat diet [156]. This effect was mediated by the altered expression in adult female offspring of 642 pancreatic islet genes, belonging to 13 functional clusters, including cation and ATP binding, and cytoskeleton and intracellular transport. Fullston et al. demonstrated the presence of altered global methylation in mature sperm and abnormal testis transcription of male mice consuming a high-fat diet, with metabolic disturbances in the next generations [157]. In addition to experimental data on animal models, very interesting data about the role played by diet on the epigenetic modifications and on the consequent transgenerational effects are available in human as well. During the winter of 1944–45 of World War II, in the Netherlands, as a reprisal against the activity of the Dutch government-in-exile aimed to disrupt the transport of German reinforcements and troops, the Germans banned all food and fuel transports to Netherlands, inducing a severe famine, with the official daily rations for the general adult population decreasing gradually from about 1800 calories (December 1943) to below 800 calories (April 1945). The situation improved in a very short time after the liberation of the Netherlands on May 1945, with the rations raising up to over 2000 calories a day by June 1945 [158]. The famine caused a severe mortality in the population of Amsterdam, but nevertheless, several babies were conceived and birthed during that period. Several decades later, a number of studies investigated the health status of people born in Amsterdam during the famine period in order to shed light on the effects of malnutrition on the health of the offspring in adult life. In a first time, these studies evidenced an association with chronic diseases in adult life in the offspring (coronary heart disease, atherogenic lipid profile, obesity, raised levels of plasma fibrinogen, and decreased levels of factor VII), strongly related to the timing in gestation of exposure to famine [158, 159]. However, by analyzing the results of these epidemiological studies with the aid of molecular tools, it has become clear that the Dutch famine families study has provided the first direct evidence for epigenetic programming through prenatal famine exposure. In fact, it was clearly demonstrated that periconceptional exposure to famine produced an under-methylation (likely related to a deficiency in methyl donors) in the differentially methylated region (DMR) of the maternally imprinted IGF2 gene [160], suggesting that early undernutrition can cause epigenetic changes persisting throughout life. On the other hand, there was no variation in IGF2 methylation status in individuals exposed to famine in later gestation. Further studies evidenced that persistent changes in DNA methylation represent a common consequence of prenatal famine exposure and that they can be affected by the sex of the exposed individual and the gestational timing of the exposure [161]. More recently, it has been demonstrated that prenatal malnutrition-associated DMRs (P-DMRs) mostly occur in regulatory regions of genes showing differential expression during early development [162].

All the above-reported studies suggest that undernutrition plays a direct effect on the fetus during the early phases of development. However, why should we exclude that the target of the famine could be represented also by gametes other than by embryo? Several evidences strongly suggest a role played by paternal diet on the healthy status of children, as confirmed by the above-described results on animal models. In human, the link between grandparental nutrition and grandchild’s growth was at first examined by Bygren et al., who demonstrated that a surfeit of food in the environment when the paternal grandfather was a boy was related to a shortening of the proband survival [163]. The same group subsequently demonstrated that a limited availability of food during the father’s prepuberal age was related to a low cardiovascular disease mortality of the proband, while paternal grandfather exposure to a surfeit of food during the same period was related to increased diabetes mortality in the proband [164], suggesting epigenetic inheritance as a strong candidate for these phenomena [165]. Soubry et al. demonstrated the presence of alterations in the methylation status at multiple imprint regulatory regions in children with obese parents, suggesting a preconceptional influence of parental lifestyle and nutrition on the programming of imprint marks during gametogenesis [166]. In particular, the significant association between paternal obesity and altered methylation in the offspring suggests the susceptibility of the developing sperm to environmental insults. Very recently, the evidence of a role played by changes in paternal grandmothers’ early food supply on the risk of cardiovascular mortality of the female grandchildren have also suggested an X-linked epigenetic inheritance via spermatozoa [167]. However, it has also been stressed that a true transgenerational inheritance in response to diet should be examined on the third generation, which represents the actual first “unexposed” one, being the first filial generation directly exposed to the maternal diet, and deriving the second from gametes exposed in utero [168]. Despite these limitations, the possible presence of transgenerational effects related to the epigenetic effect of paternal diet remains a very interesting topic.

The four windows of epigenetic susceptibility

A crucial question concerning the role played by environmental agents in the epigenetic modifications of the male gamete is the following: when are the effects of such exposures transferred to the male gamete? Soubry et al. identified four potential windows of susceptibility during the development of the paternal germ line and zygote [145]. The first window is represented by paternal embryonic development, when PGC undergo genome-wide epigenetic erasure during migration to the genital ridge. Defects in this process, as well as in the maintenance of some protected regions, could be caused by internal or external factors during early development. The second window is represented by paternal prepuberty, since in this period, de novo methylation at imprinted gene loci occurs. The third window can be identified in the period in which spermatogenesis, and in particular the development from spermatogonium to spermatocytes, occurs, since methylation patterns are established during this time. This window appears to be a very important one, representing the reproductive period of the subject, when a careful evaluation of his lifestyle, with prevention of environmental stressors, could be used as a preventive strategy. Finally, the fourth window is represented by the periconception period and the zygote stage, when histone retention in certain genes could represent a potential mechanism for inheritance of environmentally induced epigenetic marks. While studies investigating the effect of environmental agents in the latter three windows are currently carried out, mostly on animal models but also in humans, it is more difficult to investigate epigenetic modifications in PGCs. However, the recently reported discovery that amniotic fluid stem cells (AFSCs) share a number of features with PGCs [169] provides a novel cellular model for the study of the effect played by environmental agents in altering the epigenetic processes occurring in these cells, opening new scenarios in this field of study.

Finally, it must be stressed that aging represents another possible risk factor for epigenetic modifications increasing the risk of neuropsychiatric disorders such as autism and schizophrenia. Jenkins et al. have recently identified 139 regions significantly and consistently hypomethylated and 8 regions significantly hypermethylated with aging, with a total of 117 genes involved [170]. These authors evidenced that a portion of the age-related changes in sperm DNA methylation involves genes previously associated with schizophrenia and bipolar disorder.